Design And Implementation Of A Traffic Sign Recognition System Based On Android Platform

In modern life of vehicle travel, transport controls more stringent, t, all kinds oftraffic signs to let a person dazzling, make people always worry about the trafficviolation, the drivers must memorize traffic signs to convenient travel, visible trafficsigns in actual traffic occupies the important position that does not allow to ignore, hastips and guidance to the vehicle driver.With the rapid development of Android technology, the popularization andapplication of Android applications, more and more features in the Android systemimplementation, the realization of the "intelligent transportation system" also toAndroid mobile devices. Traffic signs everywhere, but it is difficult to traffic signrecognition. In order to realize the convenient and fast identification of traffic signs,this paper puts forward design has realized a traffic sign recognition system based onAndroid platform.In this paper, design and implementation of application systems in order to takeinto account the problem of phones running time and efficiency, and ultimately selectthe server and client mode is divided into the server PC, and Android clients. Mainlyrun on the server side image processing algorithms, image sent to the mobile client forprocessing, given the results and return to the mobile client. Mobile client Androidenvironment is mainly written in a user interface program, users log in to theapplication via mobile phone, send the image through the application to the server, andview the results.Mobile client to achieve a user login functionality, user interface after landingincluding pictures, choose local image, upload images, view the results of otherfunctions. Server-side implementation of image processing, the phone sends the imageto be recognized over the data acquisition, image acquisition after pretreatment, andextract eigenvalues image classifier training based on the eigenvalues, the final resultis given recognition, send back to the client application.In this traffic sign recognition, traffic sign recognition is the most important partof image segmentation. Through the analysis of traffic signs on the appearance ofsignificant features, using all kinds of traffic signs with different shapes, such as circle,triangle, due to the traffic signs also has a big difference on color, mainly including the base of color is red, yellow, blue, so traffic sign recognition may be regarded as aspecial kind of image recognition process. Carrying traffic sign recognition mainly forimage segmentation, you first need to color space R, G, B values for each domain wereflat, handling key part of the process is to set an appropriate threshold. We use the mostdirect and easy way is to continuously change the field values, test results, and thenchange the field values, and so forth to determine an optimal adjustment of the finalsolution, which is the key to the optimal threshold segmentation。But for the trafficsigns in the ban marks and to give way, as are red flags, you can extract0<R <imagesection255range. Of course, taking into account the effect of light and other externalfactors, the actual field values to be taken into account G, Part B setting. Of course, inconsidering the affected by external factors such as illumination, the actual domainvalues to be taken for consideration of G, setting of part B.Sometimes, only accordingto the size of the R, G, B to extract the target is very difficult, because of correlation isrelated to the R, G, B is too big, also called a R, G, B in a certain weight change, othervariables will change accordingly. We put the image from RGB color space conversionto ihs color space, ihs space image of each component on the first pass is small, so thefinal choice on ihs space division, respectively to extract specific areas. In the processof system test, through its own Android photographs of a set of traffic signs, whichmainly includes the ban and to give way sign, in the process of classifying datatraining recognition, experimental results reach the expected effect.This paper designed and implemented a traffic sign recognition system based onthe Android platform, designed from the beginning to complete a variety of vision. Thesystem can provide a smart phone traffic sign recognition, the user can unknown trafficsigns, pictures and sent to the server through identification, which gives the user thetrip has brought great convenience, with great practical value.